54results about How to "Determine accuracy" patented technology

The invention discloses a field-based method for feeding back text error correction after speech recognition, and belongs to the speech recognition field. Text sentences after speech recognition are analyzed based on errors of speed pauses of Chinese sentence structures. The method is characterized by detecting whether structures before and after a sentence separator meet the sentence pattern rules of Chinese language, finding pause errors, calculating and dividing sentences based on phoneme string similarity and converting into pinyin, converting pinyin into phoneme strings according to a phoneme table, finding sentences corresponding to the phoneme strings similar to the strings in a corpus, establishing a body based on a body controlled word query module through the controlled word table of the field, correcting the errors related to the field in the text after speech recognition through the body, outputting the matching result by a feedback module, and adding the correct identification result selected by a user and the original phoneme strings in the corpus. According to the method and system, the originally correct result of speech recognition may not be affected, and the speech recognition accuracy can be better determined through a body and feedback mechanism.

A discrimination sensor includes a light emitting device individually emitting sensing light beams onto a surface of an object, such as a bill and having a sensing area that is wide in a direction perpendicular to a scanning direction; and a light detecting device assuring a light detecting area that is wide in a direction perpendicular to the scanning direction and detecting light coming from a surface structure of the bill when the sensing light is emitted. The light emitting device and the light detecting device are integrated with each other.

A method of analyzing and practicing golf strokes, such as putting strokes, is provided. A surface is provided, wherein a golfer can practice one or more strokes in an environment simulating components of a golf course, such as a green. A club movement detector embedded in the surface and comprising a plurality of photodetectors and associated circuitry determines the position and velocity of the club face A head movement sensor detects rotational movement of the golfer's head during the golf stroke. An eye movement sensor detects movement of the eyes of the golfer during the golf stroke. A processor connected to the club movement sensor, eye movement sensor, and head movements sensor gathers movement data during the golf stroke and produces measurements corresponding to the golf stroke, such as eye, head, and club movement of the golfer. The invention includes measuring head movement data and wirelessly transmitting same for remote analysis during a golf stroke. Club head movement is sensed by an infrared ranger placed along a line perpendicular to the face of the clubhead. Eye, head, and club movement data can be combined into one or more serialized data streams, and transmitted wirelessly over one or more RF channels for remote reception and processing. Finger, wrist, and elbow movements can also be measured and analyzed during a golf stroke or a free throw setting.

The embodiment of the invention relates to the field of robots, and discloses a control method and a robot. The control method comprises the steps that a first dynamic model of the robot is obtained,wherein the first dynamic model is used for showing the relationship between the movement parameters and the feed-forward torque of a joint, and the parameters of the first dynamic model are known; the reference movement parameters of the joint are substituted into the first dynamic model, and the feed-forward torque of the joint is obtained through calculation, wherein the movement parameters comprise the joint angle, the joint angular velocity and the joint angular acceleration; according to the feed-forward torque of the joint, the current joint torque of the joint is determined; accordingto the current joint torque of the joint, the robot is controlled to move. By means of the embodiment, the problem that the robot shakes at the power-on moment and in the moving process is alleviated,and the control precision of a robot system is improved.

The present invention relates to a system and method to facilitate data mining applications and automated evaluation of models for continuous variable data. In one aspect, a system is provided that facilitates decision tree learning. The system includes a learning component that generates non-standardized data that relates to a split in a decision tree and a scoring component that scores the split as if the non-standardized data at a subset of leaves of the decision tree had been shifted and / or scaled. A modification component can also be provided for a respective candidate split score on the decision tree, wherein the above data or data subset can be modified by shifting and / or scaling the data and a new score is computed on the modified data. Furthermore, an optimization component can be provided that analyzes the data and determines whether to treat the data as if it was: (1) shifted, (2) scaled, or (3) shifted and scaled.

The invention provides an operation sample box on the basis of an RF (Radio Frequency) technology, which can enable registration information not to be wrong and not to be omitted and enable a specimen tag not to be stained and identified by mistake in the submitting and packaging process of a specimen, and a submission method thereof. To fulfill the aim, the invention adopts the technical scheme that the operation sample box comprises a box body, a cover body, a touch screen display, an identification module, a power supply module, a data storage module, a data comparison module, a controller, an electronic lock for locking the box body and the cover body, a wireless communication module and an RFID (Radio Frequency Identification Device) electronic tag.

The invention provides a marine crane controller based on an input shaping technology. The marine crane controller based on the input shaping technology is characterized in that firstly, a marine crane mathematic model based on Lagrange dynamical equations is analyzed and resolved, then according to an error of practical system modelling, a zero vibration and derivative (ZVD) input shaper is used for controlling a marine crane; and finally, simulation verification is carried out on the designed marine crane by utilizing the ZVD input shaper.

The invention discloses a carbonized combustible pyrolysis kinetic parameter calculation method based on a single-peak pyrolysis curve. The method comprises the steps: calculating the activation energy value E of combustible single-peak pyrolysis through an equal conversion rate method FWO method; assuming a plurality of reaction mechanisms, calculating an activation energy value E through a CR method, comparing the activation energy value E with a value obtained through an FWO method, determining a reaction mechanism function f(alpha) of unimodal pyrolysis of the combustible, and calculatinga pre-exponential factor A; determining a value range of solving parameters; and optimizing the pre-exponential factor A, the activation energy value E and the carbon generation rate v through a particle swarm artificial intelligence (PSO) optimization algorithm, and outputting a calculation result in the case of an optimal solution. According to the method, the difficulty of solving the unimodalpyrolysis high-order partial differential equation of the combustible is reduced; the calculation precision is high; the optimization efficiency is high; the application range is wide, the method canbe applied to single-peak pyrolysis solution of multiple groups of heating rates and multiple combustibles at the same time, and the experiment result is effectively predicted.

The invention provides a system and a method for measuring surface roughness in an online manner on the basis of laser light scattering processes. The method includes carrying out oblique incidence onlaser beams to the surfaces of measured objects at set angles; photographing scattered images on acquisition screens in reflection directions of the laser beams and extracting at least one of three feature parameters including a scattering feature parameter, a bright point ratio and a bright point gray ratio of each scattered image; computing roughness numerical values by the aid of relation curves between the feature parameters and the surface roughness. The heights of the acquisition screens are consistent with the heights of laser devices. Curve fitting is carried out on feature parametersof scattered images of standard roughness sample pieces and known roughness numerical values to obtain the relation curves. The system and the method have the advantages that the surface roughness ofmachined workpieces can be measured in the online manner by the aid of the method, accordingly, the quality of products can be guaranteed, and the productivity can be improved; the system and the method are combined with transparent window measurement modes, and accordingly the surface roughness (Ra ranging from 0.025 micrometer to 0.8 micrometer) of ground metal can be measured in the online manner under cooling liquid machining conditions.

The invention relates to a detector array target calibration system and method. The system comprises a light spot acquisition device, a first spectroscope, a second spectroscope, a power meter, a calibration detector module, and a to-be-calibrated detector array target, wherein the calibration detector module and the to-be-calibrated detector array target are in the same distance to the first spectroscope; light spots output by the light spot acquisition device are transmitted to the calibration detector module and the detector array target respectively after passing through the first spectroscope, the second spectroscope arranged on an optical path is used for emitting a beam of light to the power meter; and a first two-dimensional scanning platform and a second two-dimensional scanning platform enable the corresponding calibration detector module and the detector array target to perform full-coverage scanning of a set path on a plane directly facing the light spots. According to thesystem in the invention, the calibration detector module is matched with the power meter so as to obtain the real power density of a calibration light source output by the light spot acquisition device, and the two-dimensional scanning platforms are arranged, so that the straightness and running precision during full-coverage scanning of the calibration detector module and the detector array target on the set path are ensured.

The invention provides a method for changing an image local area, and belongs to the technical field of image processing. The method comprises: acquiring a first image and a pixel vector, and determining a local area in the first image; and mapping in the local area by using the pixel vector to generate a local area with a pixel feature corresponding to the pixel vector. According to the method, the automobile picture with the real automobile body color can be generated, the diversity of advertisement materials is improved, the color changing accuracy of the model is improved, the automobile body segmentation capability is further achieved, and the color changing uniformity is improved in the color changing process of the model.

The embodiment of the invention discloses an image classification method and device, computer equipment and a medium, and belongs to the field of image processing. The image classification method comprises the steps of calling a first classification sub-model in a classification model, and performing classification processing on a target image to obtain a first prediction result of the target image; calling a second classification sub-model in the classification model, and performing classification processing on the target image to obtain a second prediction result of the target image; callinga third classification sub-model in the classification model, performing classification processing on the target image to obtain a third prediction result of the target image, training sample data ofthe third classification sub-model comprising a sample image and a result obtained by statistics according to a plurality of annotation results; and outputting the third prediction result and the classification difficulty of the target image, determining the classification difficulty according to the first prediction result and the second prediction result, indicating the accuracy of the model output result through the classification difficulty, and avoiding misleading of the model to the user.

Electrophoretic displays use motion of particles through a fluid to generate images. This invention describes the use of concerted motion of particles together with the solvent to generate optical switching.

An embodiment of the invention provides a method and equipment for constructing a strength criterion material parameter of a rock sample. The method comprises the following steps: obtaining the relation between the material parameter and confining pressure in a curve fitting manner according to the rock sample and a bedding angle; obtaining the relation between the material parameter and the bedding angle in a curve fitting manner according to the rack sample and the confining pressure; obtaining the relation between the material parameter and the confining pressure and the bedding angle by using a user-defined fitting function in combination with the relation between the material parameter and the confining pressure and the relation between the material parameter and the bedding angle, and constructing a predicted value of the strength criterion material parameter of the rock sample. With the adoption of the method and the equipment for constructing the strength criterion material parameter of the rock sample in the embodiment, the predicted value of the strength criterion material parameter of the rock sample can be constructed effectively, accuracy of the predicted value of thestrength criterion material parameter of the rock sample can be further determined, and an improved H-B criterion is obtained on the basis.

The embodiment of the invention discloses a data quality determining method comprising steps of: receiving the acquisition data of a acquisition system, the first data of respective interfaces of a network, and the second data of a drive test system; and generating, based on the first data and the second data, the complete user information of the respective interfaces; generating user information to be verified based on the acquisition data; and determining a data quality index of the acquisition system based on the complete user information and the user information to be verified. The embodiment of the invention also discloses a data quality determining device.

The invention discloses a bridge and a plucked musical instrument. The bridge comprises a body and a playing signal detection assembly. The playing signal detection assembly is arranged on the body. When the bridge is installed on a plucked musical instrument, the playing signal detection assembly is corresponding to strings of plucked musical instruments, and is used for acquiring playing signalsgenerated when strings of the plucked musical instrument are played. The playing signals comprise plucked signals and pressing signals, and after a controller on the plucked musical instrument receives the playing signals acquired by each bridge, the playing signals are converted into musical instrument digital interface MIDI signals and output. According to the technical scheme, a bridge provided with a playing signal detection assembly is arranged on a plucked musical instrument, a playing signal generated when strings of a plucked musical instrument are played is obtained through the playing signal detection assembly. The corresponding playing signal is converted into an MIDI signal, the musical scale and playing information of playing of a player are accurately evaluated, and the playing accuracy of the player is determined.

The invention provides a high-dynamic-range (HDR) image quality evaluation method based on tensor domain curvature analysis. Tensor decomposition in the method can maintain most color information of an HDR image, and curvature analysis can extract structure information of the HDR image; an effective feature set is constructed through combination of tensor decomposition and curvature analysis, andis used to characterize different distortion degrees of HDR image blocks; then local features of the HDR image are extracted from a label matrix reconstructed and obtained by utilizing the feature setand a sparse dictionary with labels; and finally, predicted quality of the image is obtained through aggregating the local features and global features of the HDR image. The method of the invention is tested on two public databases; and experiment results show that performance indexes thereof are all superior to those of other non-reference metrics, and it means that consistency of the method with human visual perception is higher.

The invention discloses an electrochemical judgment method of fire-damaged concrete reinforcement damage. The method comprises the steps as follows: (1) preparation of a test piece; (2) simulation of a fire damage test; (3) electrochemical detection and data acquisition; (4) data analysis and damage evaluation. Test study is conducted on fire-damaged concrete reinforcement damage according to the basic principle of electrochemical detection of reinforcement corrosion on the basis of three electrochemical elements, an electrochemical criterion of the fire-damaged concrete reinforcement damage is established, and then, diagnostic assessment is performed on actual fire-damaged concrete structures by use of the criterion and an auxiliary analysis model.

The invention discloses a calibrating device for grounding connection parameter testing systems. The calibrating device comprises a case, an interference signal simulation module display (3), a loop impedance simulation module display (5), an impedance standard display (7) and an adjustable frequency standard signal source display (14) which are mounted on the case are respectively connected with an interference signal simulation module (2), a loop impedance simulation module (4), an impedance standard (6) and an adjustable frequency standard signal source (13) in the case, a current loop input interface terminal (1), a current loop output interface terminal (9), a first calibrating device sampled voltage interface terminal (11) and a calibrating device sampled voltage interface terminal (12) are arranged on the case, and thereby the calibrating device is formed. The impedance standard is embedded in the calibrating device, real-time frequency, current and voltage values are acquired, and a standard impedance value is compared with the impedance value displayed by a tested system in order to determine the accuracy of the tested system.

The invention discloses a renewable paper product self-service recovery system and a method, a self-service machine is used for controlling the opening of an inlet of a recovery channel after being triggered by a user, and a paper product conveying device is used for conveying a renewable paper product to an image acquisition area after detecting that the renewable paper product is placed at the inlet of the recovery channel, a camera is used for collecting a multi-view appearance image of the paper product in the image collection area, the state parameter obtaining module is used for calculating the state parameter of the paper product based on the multi-view appearance image, and a recoverability calculation module is used for calculating the recoverability of the paper product accordingto the state parameter. And the self-service machine is also used for carrying out user confirmation on the detected state parameters, issuing a recycling reward matched with the recyclable degree tothe user after receiving state parameter confirmation information fed back by the user, and triggering the recyclable paper product to be conveyed to a paper product recycling device. According to the system, a complete paper product self-service recovery process is established, so that a user can realize self-service paper product recovery without the help of others.

The invention relates to the technical field of biology, in particular to a marker, reagent or kit for detecting whether COVID-19 disease is cured. A plasma protein marker combination for diagnosing curing of a COVID-19 patient comprises a serine protease inhibitor 3, lymphocyte cytoplasmic protein 1, a peptidase inhibitor 16 and fetuin B. The results show that the area under the curve (AUC) valueof a receiver operating curve (ROC) of the four protein combinations predicting the cure of the COVID-19 patients is 0.947, the true positive rate of predicted cure is 82%, and the false positive rate is 18%. In addition, an enzyme-linked immunosorbent assay (ELISA) experiment finds that the protein concentration is significantly different in the plasma of the COVID-19 patient and a cured patient, which further confirms the accuracy of the combination for predicting the cure of the COVID-19 patient.

The invention provides a deepwater submersible vehicle charging platform capable of lifting and aligning. The charging platform comprises: a telescopic cover plate which is contracted and expanded through a cover plate control system; a charging device which comprises a rechargeable battery pack, a charging jack, a contact sensor and a clamping groove; the charging jack, the contact sensor and theclamping groove form a square structure, a drying device, a separable base plate and a charging socket are arranged in the charging jack, and a plurality of proximity sensors are arranged on the charging device; a lifting system arranged at the bottom of the charging device and used for controlling lifting of the charging device, wherein the bottom end of the lifting system is connected with a rigid bottom column, and the charging platform is fixed in a deep sea seabed through the rigid bottom column; and a positioning system which adopts a GPS + LBS positioning system, is connected with thesignal receiving device and is used for positioning a deep-sea submersible vehicle. The charging platform can ensure that the deep-sea space station can safely perform long-period and all-weather deep-sea operation, deal with many emergency situations, and save a large amount of time and energy for the deep-sea submersible vehicle to return to a wharf.

The invention discloses a method for detecting cross-border e-commerce items, cross-border express items or cross-border mail items. The cross-border e-commerce item, cross-border express item or cross-border mail item detection method based on the block chain comprises the following steps: receiving scanning information related to a cross-border e-commerce item, a cross-border express item or a cross-border mail item from a preset scanning detection device, determining cross-border goods information related to cross-border e-commerce goods, cross-border express items or cross-border mail items according to the scanning information; acquiring cross-border cargo item data related to cross-border e-commerce items, cross-border express items or cross-border mail items from a block chain system; and determining the legality of the cross-border e-commerce item, the cross-border express item or the cross-border mail item according to the cross-border goods item data and the cross-border goods item information.

The invention discloses a group exercise, aerobics and dance teaching demonstration device which comprises a platform, an assistant platform, a photographing device, a projection device, a protectionframe, an assistant collection device and a control console. A front image or a back image of a dancer is acquired through the photographing device, a learner can visually watch front action or back action of a coach at the same time and watch the back action or the front action on a projection screen, in this way, good teaching demonstration effect and accurate and quick learning are realized; stepping position information of an excellent athlete and a demonstrator is compared with corresponding note information in music playing content information, trails formed by the stepping position information of the excellent athlete and the demonstrator are compared, front image information or back image information of the excellent athlete and the demonstrator is compared, then accuracy of movement posture of the demonstrator to the music playing content information is determined, so that accuracy of the movement posture of the demonstrator and music matching degree, force applying sequence and momentum can be judged.

The present invention relates to a system and method to facilitate data mining applications and automated evaluation of models for continuous variable data. In one aspect, a system is provided that facilitates decision tree learning. The system includes a learning component that generates non-standardized data that relates to a split in a decision tree and a scoring component that scores the split as if the non-standardized data at a subset of leaves of the decision tree had been shifted and / or scaled. A modification component can also be provided for a respective candidate split score on the decision tree, wherein the above data or data subset can be modified by shifting and / or scaling the data and a new score is computed on the modified data. Furthermore, an optimization component can be provided that analyzes the data and determines whether to treat the data as if it was: (1) shifted, (2) scaled, or (3) shifted and scaled.

The invention discloses a fire rescue closed-loop control system and method based on a fire-fighting robot. On the fire-fighting robot body where the main operating system is placed, a drone docking point for docking the drone is provided on the fire-fighting robot body, and a fire-fighting bomb delivery box is provided in the fire-fighting robot body, and the bottom of the fire-fighting bomb delivery box is connected There is a fire extinguishing bomb conveying compartment arranged on an incline, and the main controller is connected with a valve for controlling the communication between the fire extinguishing bomb conveying compartment and the fire extinguishing bomb drop box. This structure realizes real closed-loop control, and finally improves detection and fire extinguishing efficiency.

An information storage system employs rotatable disks with embedded tracks for precise control of writing and reading operations. Advantage is taken of the reference tracks on one of the disks to also control laser spot positioning on the other side of the disk. The edges of the reference tracks are employed to assure accurate positioning and the presence of multiple information tracks between adjacent reference tracks.